The Research On Ultrasonic Defecting Method Based On Topological Neural Network

Oil is an important strategic material of the country and it occupies the irreplaceable position in industrial development.China has established a perfect oil reserve base and has nearly 7000 oil storage tanks.Affected by the environment and the working conditions,oil storage tanks are easy to be corroded,especially the bottom of the oil storage tanks.So,to find a more efficient method of the bottom plate defects detection is of vital significance.Firstly,the dispersion characteristics of the Lamb wave are studied and the dispersion curves of Lamb wave in steel plate are plot.Also,the principle of ultrasonic flaw detection is introduced.Secondly,the propagation of Lamb wave in the steel plate is simulated in ABAQUS software.The localization principle of ellipse is used to find the defect location.Different types of defects and the layout of sensors are also researched.Thirdly,the topology of neural network is trained to identify the type of defect.Finally,an experimental system of ultrasonic nondestructive testing based on LabVIEW is build up to verify the accuracy of the theory and numerical simulation.As a result,the dispersion characteristics of the Lamb wave are observed.The localization principle of ellipse can identify the location of different defects effectively.The recognition effect is better when the sensors in matrix arrangement than the sensors in linear arrangement.The trained topology of neural network can identify the type of defect,the recognition rate is up to 93.33%.The experimental system of ultrasonic nondestructive testing based on LabVIEW can also identify the defect of the steel plate,the results of the experiment match well with theory and numerical simulation results.